thanks jack. I did as you told me
reminding my condition composition
cue(200ms)-blank(100)-target(until subj. resp)-target2(until subj resp). TR=3000ms
i set for each of the 6 conditions: EV1 (onset:target, duration: RT1, str:1), EV2 (onset:target2, duration: RT2, str:1)
then 2 further EV including erroneous trials (one for target, one for target2), a further EV for null condition
for each subject/session I have from 0 (most of the time) to 4-5 errors (out of 120 valid trials). so things didn't change after inserting erroneous EV.
but what about the cue events?.
since they are so short(200ms) and so close to target(100ms), i imagine that their activity cannot be reasonably calculated (so I won't create contrast with him), but ...is it correct to not model it at all?.
should I
- leave as now
- create a third EV (onset:cue, duration: 0.2, str:1)
- enlarge EV1 (onset:cue, duration: 0.3 + RT1, str:1)
if I had to remove temporal derivation, should I then make a slice timing correction (slices NOT interleaved)
sorry for all these questions...
thanks in advance
cheers
Alberto
|